Azure Ai Transcription Py

Azure AI Transcription SDK for Python. Use for real-time and batch speech-to-text transcription with timestamps and diarization. Triggers: "transcription", "speech to text", "Azure AI Transcription", "TranscriptionClient".

Security Scan

The skill's instructions match a legitimate Azure transcription client, but the metadata omits the environment credentials and install step the runtime docs clearly require, which is an incoherence you should understand before installing.

Purpose & Capability

The name, description, and SKILL.md consistently describe an Azure speech-to-text client (batch and streaming). The examples reference azure-ai-transcription and Azure endpoints, which is coherent with the stated purpose. However, the registry metadata declares no required environment variables or credentials while the runtime docs require TRANSCRIPTION_ENDPOINT and TRANSCRIPTION_KEY — this mismatch is unexpected.

Instruction Scope

SKILL.md instructions are narrowly scoped to installing the client, configuring an endpoint/key, and calling the TranscriptionClient for batch or real-time transcription. The only file access shown is audio files (local or blob URLs) needed for transcription. There are no instructions to read unrelated system files, credentials, or send data to endpoints outside Azure/storage examples.

Install Mechanism

The skill is instruction-only (no install spec), but the SKILL.md explicitly tells users to run `pip install azure-ai-transcription`. That means installing a PyPI package at runtime (network fetch, code execution). This is expected for a Python SDK but is a non-trivial step not reflected in metadata — consider verifying the package origin and integrity before installing.

Credentials

The runtime docs require TRANSCRIPTION_ENDPOINT and TRANSCRIPTION_KEY (Azure endpoint and subscription key). Those credentials are appropriate for the described service, but the skill metadata declares no required env vars or primary credential. The mismatch is a red flag because credentials are necessary for operation yet not declared in the manifest; verify where/how the key will be provided and that it has least-privilege permissions.

Persistence & Privilege

The skill does not request persistent presence (always: false) and is user-invocable. There is no indication it modifies other skills or system-wide settings. Autonomous invocation is allowed by default but not combined with other high-risk signals here.

What to consider before installing

This SKILL.md describes a legitimate Azure transcription client, but the package and credentials handling need verification before installing. Before you install or run it: 1) Confirm the pip package name (azure-ai-transcription) on PyPI and that it is the official Azure SDK or a trusted publisher; prefer pinned versions. 2) Don't paste your TRANSCRIPTION_KEY into public places; use least-privilege keys and consider managed identity or safer auth if available. 3) Be aware pip install fetches and runs code from the network — audit the package or review its source first. 4) Ask the skill author or registry maintainer to declare required environment variables (TRANSCRIPTION_ENDPOINT, TRANSCRIPTION_KEY) in the metadata so the manifest matches runtime requirements. If you cannot verify the package origin or the skill metadata remains inconsistent, treat it as risky and avoid installing.

Azure AI Transcription SDK for Python

Client library for Azure AI Transcription (speech-to-text) with real-time and batch transcription.

Installation

pip install azure-ai-transcription

Environment Variables

TRANSCRIPTION_ENDPOINT=https://<resource>.cognitiveservices.azure.com
TRANSCRIPTION_KEY=<your-key>

Authentication

Use subscription key authentication (DefaultAzureCredential is not supported for this client):

import os
from azure.ai.transcription import TranscriptionClient

client = TranscriptionClient(
    endpoint=os.environ["TRANSCRIPTION_ENDPOINT"],
    credential=os.environ["TRANSCRIPTION_KEY"]
)

Transcription (Batch)

job = client.begin_transcription(
    name="meeting-transcription",
    locale="en-US",
    content_urls=["https://<storage>/audio.wav"],
    diarization_enabled=True
)
result = job.result()
print(result.status)

Transcription (Real-time)

stream = client.begin_stream_transcription(locale="en-US")
stream.send_audio_file("audio.wav")
for event in stream:
    print(event.text)

Best Practices

1. Enable diarization when multiple speakers are present
2. Use batch transcription for long files stored in blob storage
3. Capture timestamps for subtitle generation
4. Specify language to improve recognition accuracy
5. Handle streaming backpressure for real-time transcription
6. Close transcription sessions when complete